Notch3 destabilizes Regulatory T cells to Drive Autoimmune Neuroinflammation in Multiple Sclerosis

Treg cells of MS patients have increased Notch3 expression

Analysis of peripheral blood mononuclear cells (PBMCs) from patients with actively relapsing MS revealed an increase in Notch3 expression, but not of other Notch receptors, specifically in Treg cells (Fig. 1A-B; Table S1), but not in CD4⁺ T conventional (Tconv) (Fig. 1C). Treg cell Notch3 expression was decreased in patients with inactive MS compared to those with active disease and to Healthy controls (Fig. 1D,E). Analysis of cerebrospinal fluid (CSF) T cells of patients with active MS revealed a further increase in Notch3 expression in Treg cells compared to their matched counterparts in the peripheral blood (Fig. 1F). Notch3 expression was also increased but to a lesser extent in Tconv cells in the CSF of MS patients, while Notch3 expression was unchanged in the peripheral blood and CSF T cell subsets of control subjects (Fig. 1G). The frequencies of Tconv cells expressing T helper (Th) cell cytokines implicated in MS pathogenesis, including IL-17⁺ and IL-17⁺IFNγ^{+ 5,38}, were increased in the CSF but not PBMCs of MS patients (Fig. 1H-I). Notch3⁺ Treg cells in the CSF also expressed increased levels of IL-17, IL-22 and IFNγ, indicative of their subversion into a Th cell-like phenotype (Fig. 1JK). These results suggested a putative role for Notch3 in the pathogenesis of MS.

Notch3⁺ Treg cells promote EAE pathogenesis by augmenting CNS Th17 responses

To examine the functional role of Notch3 in MS, we analyzed Notch receptor expression in EAE induced by immunization of mice with a myelin oligodendrocyte glycoprotein peptide (MOG^35-55). In agreement with the human data, Notch3, but not other Notch receptors, was also increased in mouse CNS tissue CD4⁺ T cells including Treg (Foxp3⁺) and non-Treg (Foxp3^–) cells (Fig, 2A,B). Global deletion of Notch3 attenuated EAE disease severity (Fig. 2C,D). Furthermore, conditional deletion of Notch3 in T cells or more specifically in Treg cells using a floxed Notch3 allele (CD4^CreNotch3^Δ/Δ and Foxp3^YFPCreNotch3^Δ/Δ, respectively) similarly attenuated EAE disease severity, pointing to Treg cells as the key drivers of Notch3-mediated EAE pathology (Fig. 2C,D). Moreover, phenotyping of different immune tissues in Foxp3^YFPCreNotch3^Δ/Δ mice failed to demonstrate any spontaneous immune activation or skewing (Fig. S1A-N). In contrast, Treg cell-specific deletion of Notch1, Notch2 or Notch4 had minimal (Notch1) or no effect (Notch2, Notch4) on EAE disease outcome and cytokine production by CNS CD4 cells after EAE induction (Fig. S2A-D). These results confirmed a specific pathogenic role for Notch3⁺ Treg cells in EAE.

To further decipher the role of Notch3⁺ Treg cells in EAE pathogenesis, we carried out single-cell RNA sequencing (scRNA-seq) on CD4⁺ T cells from the CNS of MOG-immunized Foxp3^YFPCre and Foxp3^YFPCreNotch3^Δ/Δ mice. We first performed a graph-based clustering analysis using Seurat, which uncovered 12 clusters (Fig. 2E and S3C-F). These clusters included a Treg cell cluster (C0) expressing canonical Treg cell genes (e.g. Foxp3, Ikzf2, Il2ra), an effector CD4⁺ T cell cluster (C1) (Tbx21, Id2, Cxcr6), a Th17 cluster (C3) (Rorc, Ccr6, Il23r), a T follicular helper cell cluster (C4) (Icos, Pdcd1, Batf), and a mixed Treg/Th17 cluster (C7) (Foxp3, Rorc, Maf) (Fig. 2E, S3F). We performed pseudo-bulk differential gene expression analysis in C0 and C3 followed by pathway and upstream regulator analysis. Results revealed that C0 cluster of Foxp3^YFPCreNotch3^Δ/Δ mice showed increase in a Treg cell pathway (RUNX1 and Foxp3 control the development of Tregs) and decrease in a pathway involved in Treg cells destabilization (PTEN, PI3K/AKT signaling). The C0 cluster of Foxp3^YFPCre had increased MyD88, Hippo pathway and TLR signaling (Fig. S3G-I). Pseudo-bulking analysis of cluster C3 also revealed an increase of MyD88 and TLR signaling in the C3 cluster of Foxp3^YFPCre (Fig. S3K-L). Cell proportion analysis revealed that only two clusters were significantly different between the two groups: a naïve CD4 cell cluster (C11) and a Treg/Th17 cluster (C7) (FDR<0.05, see Methods). To further decipher differences in the C7 transcriptomic programs between the respective groups, we performed pseudo-bulk differential gene expression analysis followed by pathway and upstream regulator analysis (Fig. 2F, I). Results revealed that the C7 cluster of Foxp3^YFPCreNotch3^Δ/Δ mice showed decreased pathways associated with Th17 cell differentiation, including Hippo ^35,39,40 and STAT3, and increased genes targeted by Foxp3 compared to the C7 cluster of Foxp3^YFPCre group. Upstream regulator analysis performed by Ingenuity pathway analysis (IPA, Qiagen) predicted increased activation of several pathways including Notch (recombination signal-binding protein for immunoglobulin kappa J region, RBPJ) and Hippo (yin yang 1, YY1 and tyrosine 3-monooxygenase/tryptophan 5 monooxygenase activation protein gamma, YWHAG) in the C7 cluster of Foxp3^YFPCre group (Figure 2G). We further performed pseudo-bulk differential gene expression analysis between clusters C0 and C7 followed by pathway analysis (Fig. 2H and I). Results revealed that cluster C7 presents a decrease in the Treg cell canonical gene Ikzf2 and an increase in gut Treg cell-associated genes including Maf and Ffar2 ^41,42 (Fig. 2H). Pathway analysis revealed increased pathways associated with Th17 cell differentiation (Th17 activation pathway, IL-6 family signaling, STAT3 pathway) (Fig. 2I). Overall, these results suggest that Notch3 destabilized Treg cells towards a Th17 phenotype (Fig. 2F-I). Consistent with the results of the scRNA-seq study, analysis of IL-17 producing T cells frequencies in EAE revealed decreased IL-17⁺IFNγ⁺, IL-17⁺GM-CSF⁺, IL-17⁺IL-22⁺ subsets in the CNS of mice with Notch3-deficient Treg cells compared to controls (Fig, 2J-L).

Notch3 deficiency suppressed disease independently of MOG^35-55 TCR specificity

Investigation into the antigen specificity of CNS Notch3⁺ T cells revealed that the frequency of MOG^35-55 peptide-specific CD4⁺ T cells, as measured by staining with MOG^35-55 peptide-specific I-A^b tetramers, was sharply decreased in the CNS of Foxp3^YFPCreNotch3^Δ/Δ compared to Foxp3^YFPCre mice following EAE induction (Fig. S2E, F). To discriminate the respective roles of MOG^35-55 peptide-specific T cell receptors (TCR) versus Notch3 in the pathogenesis of EAE, we examined disease progression in Notch3-sufficient or -deficient mice carrying a MOG^35-55 peptide-specific TCR transgene (2D2) ⁴³. Global Notch3 deletion suppressed the incidence of spontaneous EAE in 2D2 transgenic mice (Fig. S2G). Moreover, induction of EAE and IFNγ⁺ IL17⁺ producing CNS CD4 cells was abrogated in Rag1^−/− mice that received Notch3-deficient compared to Notch3-sufficient 2D2⁺ T cells followed by immunization with MOG^35-55 peptide (Fig. S2H-L).To further decipher the decrease in MOG^35-55 peptide-specific CD4⁺ T cells at peak of disease, we analyzed the expression of the co-stimulation markers CD80 and CD86 and MHC-II in antigen presenting cells (APCs) and also the frequencies of MOG^35-55 peptide-specific CD4⁺ T cells in the draining lymph nodes at day 6 post disease induction, at which time T cell priming in EAE takes place ⁴⁴. Results revealed that APC maturation and MOG^35-55 peptide-specific CD4⁺ T cell frequencies in the draining lymph nodes were not impacted by Treg cell Notch3 deletion (Fig. S2M-P). Overall, these results suggested that Notch3 deficiency suppressed disease induction separately from its effect on MOG^35-55 peptide-specific TCR cell frequencies or T cell priming.

Notch3 is induced by the gut microbiota via Treg cell Toll-like receptors (TLRs).

Analysis of Notch3 expression in Treg and Tconv cells in different tissues at steady state revealed that they were primarily found in the gut and the CNS (Fig. S4A). Flow cytometric analysis of Notch3⁺ Treg cells revealed decreased Foxp3 mean fluorescence intensity (MFI) and increased expression of CNS homing (CCR5, CD49d, CXCR3) ^45-48 and gut homing marker (α4β7) (Fig. 3A-H). CNS CD4⁺Notch3⁺ Treg and non-Treg cells from Foxp3^YFPCre mice were highly enriched in MOG^35-55 peptide-specific tetramer staining after EAE induction (Fig. 3I-J), suggesting that Notch3 may translocate from the gut to the CNS to promote EAE pathogenicity due to their TCR specificity and the expression of CNS homing markers. CNS Notch3⁺ Treg cells, isolated after EAE induction, exhibited decreased in vitro suppressive function compared to their Notch3^– counterparts (Fig. S4B), consistent with a disease promoting function of Notch3 in CNS Treg cells.

Previous studies implicate the gut microbiota and a gut/CNS axis in EAE and in MS ^49-55. To analyze the role of the gut microbiota in Notch3 induction, we treated mice with the antibiotic ampicillin, previously reported to suppress EAE induction ⁵². Ampicillin treatment attenuated EAE in Foxp3^YFPCre mice in association with decreased CNS Notch3⁺ Treg cells, suggesting that Notch3 is induced by a gut-microbiota-dependent mechanism (Fig. S4C, D). Consistent with their gut microbiota-dependent ontogeny, Notch3⁺ Treg cells were minimally detected in the gut and CNS tissues of germ-free (GF) mice as compared to specific pathogen-free (SPF) mice but were induced upon fecal microbiota transplant of GF mice (Fig. 3K). Further analysis showed that Treg cell-specific deficiency of MyD88, and to a lesser extent the TLR2 and TLR5 but not TLR4, suppressed EAE disease severity and decreased the frequency of IFNγ⁺IL17⁺ and IL22⁺IL17⁺ CNS T cells (Fig. 3L-N and Fig S4 E and F). Disease attenuation was associated with decreased expression by CNS Treg cells of Notch3 and the gut homing integrin α4β7, consistent with the induction of Notch3⁺ Treg cells by TLR-activating gut microbial signals (Fig. 3 O and P). To confirm the gut ontogeny of CNS Notch3⁺ Treg cells in EAE, we transferred naïve CD4⁺ T cells into Rag1^−/− immunodeficient mice which were then immunized with MOG^35-55. Notch3⁺ Treg cells first appeared in the gut lamina propria at day 4 post-immunization and then appeared in the CNS at day 7 post-immunization, suggesting that Notch3⁺ Treg cells are first induced in the gut lamina propria and then migrate to the CNS following EAE induction (Fig. S5A-I). To confirm the gut to CNS translocation of Notch3⁺ Treg cells, we treated MOG^35-55 immunized Foxp3^YFPCre mice with either anti-α4β7 mAb, anti-Vla-4 mAb to block the gut or CNS homing. Results revealed that treatment with anti-α4β7 or anti-Vla-4 mAbs but not with an isotype control mAb led to decreased Notch3⁺ Treg cell frequencies in the CNS and their reciprocal increase in the gut (Fig. 3Q-R). Altogether, these results confirm that Notch3⁺ Treg cells are induced in the gut by a gut-microbiota-MyD88 dependent mechanisms. They then migrate to the CNS to promote EAE pathogenesis aided by their MOG^35-55 specificity and the expression of CNS homing markers.

Notch3 destabilizes Treg cells by Hippo-pathway-dependent mechanisms

In view of the scRNA-seq data implying the destabilization of Notch3⁺ Treg cells into Th17-like cells following the migration of the former into the CNS in EAE (Fig. 2E-H), we investigated this destabilization process by performing bulk RNA-seq analysis of CNS Treg cells isolated from Foxp3^YFPCre versus Foxp3^YFPCreNotch3^Δ/Δ mice in EAE. Results revealed increased expression of gene transcripts of several pathways associated with Th cell activation and differentiation including Th17 cells, epigenetic modulation and regulatory T cell lineage commitment (Fig. 4A, B). Notch3⁺ Treg cell destabilization into Th17 cell subsets was further confirmed by a lineage tracing approach using a Foxp3^YFPCre and a Cre-activated RFP reporter inserted into the Rosa26 locus (Foxp3^YFPCreR26^LSL-RFP) ^26,56,57. Treg cell-specific Notch3 deletion prevented Treg destabilization and the generation of ex-Treg cells (CD4⁺YFP^–RFP⁺) following EAE induction (Fig. 4C). Further analysis revealed that the generation of ex-Treg cells preferentially took place in the CNS (brain, meninges and spinal cord tissues) compared to the spleen and the gut (Fig. 4D). Notch3 expression was elevated in CNS Treg and ex-Treg cells (YFP⁺RFP⁺ and YFP^–RFP⁺, respectively) but not in Tconv cells (YFP^–RFP^–) (Fig. 4E, Fig. S5J). Further analysis of cytokine expression in ex-vivo-isolated CNS Treg and ex-Treg cells of EAE mice revealed that upward of 40% of IL-17-expressing cells localized to these populations (Fig. S5K-O). Furthermore, in vitro co-culture of Notch3⁺ versus Notch3^– CNS Treg cells of EAE mice with microglial cells pulsed with MOG^35-55 peptide revealed induction of IL-17-expression in the former but not the latter Treg cells (Fig. S5P-R). In contrast, IFNγ expression was similarly low in both populations. Finally, longitudinal analysis revealed that following peak disease at day 15 and the subsequent decline of Notch3⁺ Treg cell frequencies, Notch3⁺IL17⁺ and Notch3⁺IL-17⁺IFNγ⁺ ex-Treg cells persisted at day 30 post-EAE induction (Fig. S5S-T). These results indicated that Notch3⁺ Treg cell destabilization is a major mechanism driving the Th17 cell response in EAE.

To determine the molecular mechanisms involved in Notch3-mediated Treg cell destabilization into Th17 cells, we focused on the role in this process of the Hippo pathway, which was implicated in Treg cell destabilization in previous studies ^35,39 and by our scRNA-seq analysis (Fig. 2 F and G ). Notch3 deletion profoundly inhibited the expression of the Hippo pathway effector Yap1 in CNS Treg cells in EAE (Fig. 4F). Consistent with these results, Treg cell-specific deletion of Yap1 and Wwtr1, the latter encoding the Yap1 related effector Taz, suppressed disease severity in EAE to a similar extent as Notch3 deletion (Fig. 4 G, H). Combined Yap1 and Wwtr1 deletion did not affect the frequency of Notch3⁺ Treg cells, but greatly reduced the frequency of Notch3⁺ Tconv cells (Fig. 4I). Lineage tracing analysis using Foxp3^YFPCreR26^LSL-RFP and Foxp3^YFPCreYap1^Δ/ΔWwtr1^Δ/ΔR26^LSL-RFP mice demonstrated that Treg cell-specific deletion of Yap1 and Wwtr1 suppressed ex-Treg generation after EAE induction (Fig. 4J). It also reduced the IL-17⁺ Tconv cell subpopulations (Fig. 4 K-L). The role of the Notch3-Hippo pathway in Treg cell destabilization was also confirmed by co-culturing Notch3-sufficient, Notch3-deficient or Yap1/Taz deficient Treg cells derived from MOG^35-55-immunized mice of the respective strains with microglia derived from Notch3-sufficient (Foxp3^YFPCre) EAE mice. Notch3-deficient and Yap1/Taz-deficient Treg cells were less destabilized in this co-culture system compared to Notch3-sufficient Treg cells, as indicated by higher expression of Foxp3 in the former cells (Fig. 4M). These results indicated that Notch3⁺ Treg cells are destabilized by a Hippo pathway dependent mechanism in EAE.

Notch3-DLL1 interaction destabilizes Treg cells and promotes pro-inflammatory microglia.

To elucidate the ligand(s) involved in Notch3⁺ Treg cell-microglia interactions in EAE, we analyzed microglial expression of the different Delta and Jagged Notch ligands. The Notch delta-like ligand 1 (DLL1) was found selectively increased in microglia in EAE in both WT and Treg cell Notch3-deficient mice compared to non-immunized mice (Fig. 5A). Co-culture of CNS Treg cells with microglia isolated from EAE mice revealed that Notch3-sufficient Treg cells were destabilized, as evidenced by decreased Foxp3 expression. Additionally, this destabilization was reversed by treatment with an anti-DLL1 but not and anti-Jagged (Jag)2-antibody. In contrast, Notch3-deficient Treg cells destabilized to a lesser extent under the same culture conditions while treatment with anti-DLL1 antibody did not improve their stability (Fig. 5B). Reciprocally, Notch3-sufficient Treg cells interacted with the microglia in a DLL1-dependent manner to increase CD86 and suppress CD206 and aryl hydrocarbon receptor (AhR) expression (Fig. S6A-C). These results were confirmed in studies showing that the microglia of Foxp3^YFPCreNotch3^Δ/Δ mice subjected to EAE showed increased expression of CD206 and AhR, both markers of anti-inflammatory M2-like microglia ⁵⁸, compared to those of similarly treated Foxp3^YFPCre mice (Fig. S6D-F). RNA-seq analysis of microglia isolated from mice in EAE also revealed enrichment of several anti-inflammatory pathways in Foxp3^YFPCreNotch3^Δ/Δ compared to Foxp3^YFPCre compared to mice (Fig. S6G,H), including the LXR/RXR ⁵⁹, sphingomyelin and Retinol pathways ^60,61.

To determine the role of Notch3-DLL1 interaction in disease pathogenesis, we treated Foxp3^YFPCreR26^LSL-RFP mice with anti-DLL1 mAb as previously described ⁶². Anti-DLL1 treatment attenuated EAE severity (Fig. 5C) and prevented CNS ex-Treg cells generation (Fig. 5D). To go further, we analyzed the impact of Treg cell-specific deletion of the gene encoding the Notch receptor modifier Lunatic Fringe (Lfng), which differentially promotes Notch receptor-delta ligand interactions by glycosylating O-fucose residues on the extracellular EGF repeats of Notch receptors ⁶³. Whereas Treg cell-specific deletion of Lfng did not affect T cell development or activation (Fig. S1A-N), it recapitulated the effects of Notch3 in inhibiting EAE and suppressed various CNS Th17 cell subsets (Fig. 5 E-H). In vitro co-culture studies with microglia from EAE mice also demonstrated improved stability of the Lfng-deficient compared to Foxp3^YFPCre Treg cells (Fig. 5I). Collectively, these results indicate that Notch3-DLL1 interaction drives Treg cell destabilization in EAE.

Notch3 deletion suppresses pathogenic IFNγ⁺GM-CSF⁺ T cells by promoting a NPY1R⁺ Treg cell population.

One key gene that emerged from the analysis of RNA-seq data of Foxp3^YFPCre compared to Notch3-deficient Treg cells in EAE is Npy1r, encoding Neuropeptide Y receptor 1, which scored as the top gene hit (Fig. 4A). The NPY/NPY1R axis plays a protective role in EAE by suppressing MOG^35-55-reactive Th1 cells ⁶⁴. Moreover, polymorphisms in the human NPY gene are associated with susceptibility to MS ⁶⁵. Previous studies show the presence of Treg cells in the meninges and CNS parenchyma ^66-68. Analysis revealed that at steady state NPY1R⁺ Treg cells were selectively enriched in the CNS (brain, Spinal cord and meninges) in mice as young as 4 weeks of age and they were maintained in 1 year old mice (Fig. 6 A; Fig. S7A). Expression of NPY1R and Notch3 on CNS Treg cells was mutually exclusive, indicating they represented distinct Treg cell populations (Fig. 6B). Whereas the number of Notch3⁺ Treg cells expanded considerably following EAE induction in the CNS of Notch3-sufficient mice (Foxp3^YFPCre), those of NPY1R⁺ Treg cells were minimally affected (Fig. 6C-D). In contrast, NPY1R⁺ Treg cells markedly increased both in frequencies and in numbers in Foxp3^YFPCreNotch3^Δ/Δ mice following EAE induction (Fig. 6C).

Further immune phenotypic analysis confirmed that NPY1R⁺ and Notch3⁺ Treg cells represented two distinct populations. NPY1R⁺ Treg cells differentially expressed markers of CNS-tissue residency including CD69, CD103, CCR5 and CXCR3 ⁶⁹, consistent with the ontogeny of the NPY1R⁺ Treg cells as brain tissue-resident Treg cells (Fig. 6 E-K). In contrast, Notch3⁺ Treg cells expressed the gut homing receptor α4β7 and had lower expression levels of Foxp3 and Helios compared to NPY1R⁺ Treg cells, consistent with their relative instability (Fig. 6E-K). Furthermore, and unlike Notch3⁺ Treg cells, the frequency of CNS NPY1R⁺ Treg cells were unchanged in GF mice compared to GF mice reconstituted with fecal microbiota transplant (FMT) or SPF mice (Fig. 6L). Ultra-low input bulk RNA-seq analysis of sorted NPY1R⁺ an Notch3⁺ Treg cells revealed that NPY1R⁺ Treg cells are metabolically distinct, being enriched in lipid and tryptophan metabolic pathways (Ankrd9, Ces2c, Mecr, Cpt2) ^70-72 and having increased expression of genes involved Treg cell function (Ctnnd1, Hmox1, Ada, Ciita, Xcl1, Dgat2, Jund) ^73-83 (Fig. 6M). Pathway analysis and upstream regulator analysis confirm that NPY1R⁺ Treg cells present changes in their metabolism whereas Notch3⁺ Treg present an increase in TLR/MyD88 activation (Fig. 6 N and O).

To further elucidate the role of NPY1R⁺ versus Notch3⁺ Treg cells in the pathogenesis of EAE, we analyzed the changes in these two Treg cell subsets over time following EAE induction in WT mice (Foxp3^YFPCre). Notch3⁺ Treg cells were found expanded at day 15 and they receded at day 30 post EAE, whereas NPY1R⁺ Treg cells exhibited decreased frequencies at day 15 primarily due to the expansion of Notch3⁺ Treg cells in the face of largely unchanged NPY1R⁺ Treg cell numbers. The frequencies of NPY1R⁺ Treg cells rebounded back at day 30 post EAE in association with modestly increased cell numbers (Fig 7A, B). We further analyzed EAE induction in WT mice (Foxp3^YFPCre) and in mice with either Treg cell-specific Notch3 or NPY1R deficiency (Foxp3^YFPCreNotch3^Δ/Δ, Foxp3^YFPCreNpy1r^Δ/Δ) or combined Notch3 and NPY1R deficiency (Foxp3^YFPCreNotch3^Δ/ΔNpy1r^Δ/Δ). Foxp3^YFPCreNpy1r^Δ/Δ mice were asymptomatic at steady state and did not develop spontaneous disease when followed for up to one year (Fig S1A-N). Whereas mice with Treg cell-specific Notch3 deficiency were protected against EAE, mice with Treg cell-specific NPY1R deficiency exhibited severe disease post EAE induction that peaked earlier (d9 for NPY1R-deficient vs. d15 for Notch3-sufficient mice) with fatal outcome (Fig. 7C, D). Furthermore, combined Treg cell Notch3 and NPY1R deficiency partially reversed the protective effect of isolated Treg cell Notch3 deficiency (Fig. 7C, D, Fig S7 B). Whereas Notch3⁺ Treg cells were enriched in MOG^35-55 tetramer staining, NPY1R⁺ Treg cells had reduced staining compared to NPY1R^– Treg cells (Fig. S7C). Analysis of cytokine expression in EAE at d9 and d15 revealed that in contrast to Notch3 deficiency, which suppressed IL-17 expressing CNS Tconv cell populations at d15, Treg cell-specific NPY1R deficiency resulted in the selective expansion of IFNγ⁺ and IFNγ⁺GM-CSF⁺ Tconv cells, implicated in the pathogenesis of MS and EAE (Fig. 7E-H) ^18,84,85. The protective role of NPYR1⁺ Treg cells in EAE was further supported in studies using the NPYR1 inhibitor BIBO03304 ⁸⁶, which largely abrogated the protective effect of Treg cell-specific Notch3-deficiency. Treatment with BIBO03304 at a high dose induced lethality and severe EAE in WT (Foxp3^YFPCre) mice, consistent with results obtained with Foxp3^YFPCreNpy1r^Δ/Δ mice (Fig. S7D-G). Overall, these results established the presence of two distinct CNS Treg cell subpopulations in EAE: one gut-derived pathogenic population expressing Notch3 and another tissue resident protective population expressing NPY1R.

Lead contact:

Talal A. Chatila, M.D., M.Sc., Division of Immunology, Boston Children’s Hospital, Department of Pediatrics, Harvard Medical School, Boston, MA, USA. Email: talal.chatila@childrens.harvard.edu

Materials availability:

Any data and materials to be shared will be released via a material transfer agreement.

Data and code availability:

Bulk-RNA (GSE253097) and Sc-RNAseq (GSE253602) sequencing datasets have been deposited in the Gene Expression Omnibus. All the script have been deposited in Github (https://github.com/OlgaLanzetta/Notch3-MS)

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

METHOD DETAILS

QUANTIFICATION AND STATISTICAL ANALYSIS

Student’s two-tailed t-test, Log-rank test, one- and two-way ANOVA and repeat measures two-way ANOVA with post-test analysis.